"Efficiency" is the wrong word, cell capacity is the proper metric. I believe it was discussed in the Roadster 3.0 thread. Also the Panasonic NCR18650A cell is 3100mAh, the NCR18650G is 3600 mAh. We've heard Elon and JB say the Tesla chemistry is slightly different, but it also fits that a change from the A to the G revision of the cell would be responsible for the capacity increase. Presumably production capacity on the G cell hasn't ramped up to the point where they can supply all production yet? But yeah, I'd say a 99.8kWh pack (100kWh with slightly adjusted charge limits, or just rounding) is quite likely soon.
Some great stuff here, but this raises a few questions. We know that the S60 uses the NCR18650A at 3100mAh, but remember that the S85 uses the Panasonic NCR18650B rated for 3400mAh. So assuming that the V/i curves for the cells were completely flat, they're not but that's an ok approximation, we find that switching to NCR18650G for the S85 would result in a pack that ~90kWh not the 99.8kWh that comes from your comparison with the 3100mAh cells.
However, there are unconfirmed rumors of a Panasonic NCR18650C with a 4120mAh rated capacity at 4.35V(recharge) rather than 4.2V. If these cells are real then Tesla could create a pack on the order of ~105kW.
So I think that this really all comes down to whether the NCR18650C is a real cell or not. If it isn't then I see no reason for tesla to update the S85 for a measly ~5kWh capacity increase. However, If the 4120mAh supercell is a real thing I could definitely see something like an S105 in the near future.
As an aside. A 105kWh pack would allow for something like ~340 miles of range in the S105D configuration. This would put ~100 miles of range between the S105D and the S70D, which could be useful from a marketing perspective. So at this point I would say that an update to the capacity of the S85 capacity is perhaps slightly more likely than not, with a very large uncertainty around that assessment.
Edit: Also there's reason to think that Panasonic would be able to produce something like the NCR18650C in 2015. In 2009 Panasonic promised a 4000mAh cell to be produced in 2013 with a silicon anode that never really materialized. The NCR18650C could finally be the production of such a cell. This has the added benefit of explaining why the recharge voltage of the rumored NCR18650 is higher than 4.2 volts, as the silicon anode may have higher over-potentials than the graphite used in more standard chemistries.
Edit 2: There seems to be a good chance that I might be wrong here. It looks like there may be a discrepancy between what most people are using here to report battery capacity; some people are using the minimum cell capacity while others are using the average capacity. I look into it and see if I can correct my mistakes.
Edit 3:
The teardown from wk057 show the discharge curve for the batteries in the 85kWh and comes up with a 2.9mAh number meaning that it has to be the NCR18650A and not the NCR18650B.
This means that Bob's original thoughts are correct.